Table 1 Age estimation result comparison between different methods and features
From: Age estimation via electrocardiogram from smartwatches
Features | Mean absolute error (MAE) | |||
|---|---|---|---|---|
LSTM | Inception1D | Feedforward neural network | Transformer network | |
De-noised ECG | 3.06 | 3.03 | 2.93 | 4.71 |
PSD | 4.61 | 5.09 | 4.34 | 5.69 |
Wavelet: sym4 | 3.96 | 3.61 | 3.34 | 4.83 |
Wavelet:db4 | 3.54 | 3.29 | 3.28 | 4.11 |
Wavelet: db1 to db4 | 3.82 | 3.75 | 3.71 | 4.61 |
Wavelet: db1 to db10 | 3.15 | 3.47 | 3.21 | 4.33 |
Wavelet: sym4, db1 to db10 | 3.21 | 3.12 | 3.00 | 3.86 |
